Anatomic Pathology / MONOPREP PAP TEST CLINICAL TRIAL RESULTS
The MonoPrep Pap Test for the Detection
of Cervical Cancer and Its Precursors
Part I: Results of a Multicenter Clinical Trial
Edmund S. Cibas, MD,1 Todd A. Alonzo, PhD,2 R. Marshall Austin, MD, PhD,3
David R. Bolick, MD,4 Michael D. Glant, MD,5 Michael R. Henry, MD,6 Ann T. Moriarty, MD,7
J. Thomas Molina, MD, PhD,8 Lynda Rushing, MD,9 Sally D. Slowman, MD,5
Roosevelt Torno, MD,10 and Carol C. Eisenhut, MD5
Key Words: Papanicolaou test; Uterine cervix; Cancer; Cancer screening; MonoPrep; Squamous intraepithelial lesion; Atypical squamous cells
DOI: 10.1309/E63PQJJXWCDLWNHQ
Abstract
The MonoPrep Pap Test (MPPT; MonoGen,
Lincolnshire, IL) is a novel, liquid-based specimen
collection and processing technology for cytologic and
molecular testing. Its usefulness in the detection of
cervical cancer and its precursors was evaluated in a
multicenter, masked, adjudicated, split-sample study of
10,739 samples. After preparation of a conventional
smear, the residuum on the collection device was rinsed
into a collection vial from which an MPPT slide was
prepared. Accuracy was assessed by masked reference
interpretation by an independent pathologist. Slides
prepared by MPPT, compared with smears, yielded
statistically significant increases in relative sensitivity
for atypical squamous cells of undetermined
significance and worse, atypical squamous cells, cannot
exclude high-grade squamous intraepithelial
lesion/atypical glandular cells and worse, and lowgrade squamous intraepithelial lesion and worse. There
was no significant difference in relative specificity.
MPPT provided a 58% reduction in unsatisfactory
slides. There was no significant difference in the
presentation of endocervical/transformation zone
component or the detection of benign conditions. The
MPPT is a promising new liquid-based technology for
cervical cancer screening.
Three liquid-based Papanicolaou (Pap) tests are approved
for use in the United States as replacements for the conventional Pap smear: the ThinPrep (Cytyc, Marlborough, MA),
SurePath (BD Diagnostics-TriPath, Burlington, NC), and
MonoPrep Pap (MonoGen, Lincolnshire, IL) tests. Liquid-based
Pap tests were originally developed to facilitate the computerized
screening of Pap slides because the identification and classification of cells from optical density measurements is challenging
with conventional smears in which the thickness and overlap of
cells are poorly controlled. Cell identification and classification
by image analysis are much less problematic with the “thin
layers” produced by liquid-based preparation methods.
Aside from their suitability for automated screening, liquid-based Pap tests alone, when compared with smears, were
shown to increase the sensitivity of the Pap test in the detection of epithelial cell abnormalities. Thus, in 1996, the US
Food and Drug Administration (FDA) approved the first liquid-based Pap test, the ThinPrep, as a replacement for the
smear.1,2 In 1999, another liquid-based Pap test, the SurePath,
formerly known as AutoCyte PREP, was also approved as a
replacement for the smear.3 In the United States, the majority
of Pap tests are performed using liquid-based methods rather
than the smear.
In March 2006, the FDA approved the most recent liquidbased preparation, the MonoPrep Pap Test (MPPT). This report
summarizes some of the results of a multicenter, masked, adjudicated, split-sample study performed to test the effectiveness
of the MPPT. In this report (part I), the accuracy of the
MonoPrep is compared with that of a conventional smear as
adjudicated by a reference pathologist. Part II will review its
accuracy based on histologic correlation, and Part III will summarize data on human papillomavirus (HPV) correlation.
Am J Clin Pathol 2008;129:193-201
© American Society for Clinical Pathology
193
DOI: 10.1309/E63PQJJXWCDLWNHQ
193
193
Cibas et al / MONOPREP PAP TEST CLINICAL TRIAL RESULTS
Materials and Methods
The prospectively designed objective of this trial was to
demonstrate that the MPPT provides a statistically significant
improvement over screening with smears for the detection of
independently confirmed epithelial abnormalities at 2 diagnostic thresholds: (1) atypical squamous cells of undetermined
significance (ASC-US) and more severe lesions (ASC-US+)
and (2) low-grade squamous intraepithelial lesion (LSIL) and
more severe lesions (LSIL+). ASC-US cases were further
stratified into “low-risk ASC-US” and “high-risk ASC-US”
based on results of testing for HPV. A study protocol was
developed, sample size estimates were made, and study design
concepts were discussed with representatives of the Office of
In Vitro Device Evaluation and Safety of the FDA; their suggestions were incorporated into the final study protocol and
data analysis.
Women who underwent cervical Pap testing in the ordinary course of medical practice were eligible for inclusion in
the study. Women without a cervix were excluded. Subjects
had to be 18 years or older and able to provide written,
informed consent. Informed consent was obtained from all
subjects.
Samples were collected at sites in the United States (n =
75), South Africa (n = 11), and Venezuela (n = 2). Sites included family planning clinics, women’s health care centers, hospital-based practices, colposcopy clinics, and physician practices specializing in obstetrics and gynecology, representing a
spectrum of high- to low-prevalence populations with diverse
ethnic and racial heritage, age, and geographic location.
Institutional review board approval was obtained from all
sites. Sample collection took place between March 1 and
October 22, 2004.
All specimens were collected using the split-sample
method. First, a smear was obtained and prepared using a
plastic spatula and an endocervical cytobrush according to
standard methods.4 The residuum on the collection devices
was then rinsed in an MPPT collection vial (MonoGen), the
vial was closed, and the collection devices were discarded. A
site coordinator completed a case report form (CRF) and a
laboratory requisition form. Information captured on the CRF
included the subject’s demographic and gynecologic history
(eg, date of birth; date of last menstrual period; race/ethnicity;
previous Pap test and HPV results, if any; previous therapy or
assessments for gynecologic malignancy; use of hormone
products; and clinical observations of cervical condition at the
time of collection). The collection site placed the MonoPrep
vials and Pap smear slides in bags that were then shipped for
examination.
The specimens and corresponding CRFs were first sent to
a central laboratory (DCL Medical Laboratories, Indianapolis,
IN) where they were checked for protocol compliance. Both
194
194
Am J Clin Pathol 2008;129:193-201
DOI: 10.1309/E63PQJJXWCDLWNHQ
Pap specimens were then forwarded to 1 of 4 study laboratories (CYTO Specialty Laboratories, San Antonio, TX; DCL
Medical Laboratory, Indianapolis, IN; Pathology Services,
Cambridge, MA; Universal Diagnostic Laboratories,
Brooklyn, NY) for processing and assessment. Each of the 4
laboratories had experience in conventional and liquid-based
Pap methods and an annual volume of at least 100,000 Pap
cases. In nearly all cases, the matching Pap smear and MPPT
specimen were sent to the same laboratory. Each laboratory
had at least 2 study-designated screening cytotechnologists, at
least 1 quality control (QC) cytotechnologist, at least 1
cytopathologist, and a study coordinator. All cytotechnologists
were certified by the American Society for Clinical Pathology,
and all cytopathologists were certified in anatomic pathology,
with added qualification in cytopathology, by the American
Board of Pathology. All cytotechnologists and cytopathologists also received training in MPPT morphology by means of
a specially designed training module.
Random numbers were assigned to the cases and affixed
to all MonoPrep vials, slides, and CRFs. CRFs and vials and
slides were separated following accessioning such that subject names, initials, and other identifiers were not available or
discernible to the cytotechnologists or cytopathologists. The
codes that linked a smear to its pair-mate MonoPrep slide
were not provided to the laboratory sites. Study personnel
were instructed not to read MonoPrep and study-associated
Pap smears on the same day. They were instructed not to
compile average or comparative results, nor were they permitted to keep or refer to any personal or other log or notes
of study-related Pap smears or MonoPrep slides that they, or
any other person, had reviewed. All cytotechnologists and
cytopathologists performed their reviews under such masked
conditions. The study coordinator was not permitted to
review slides at all.
The Pap smears were accessioned, racked, stained, and
coverslipped in accordance with customary practice.4
MonoPrep slides were prepared at the study laboratory using
the MonoPrep Processor (MonoGen) ❚Image 1❚. The
MonoPrep Processor is a fully automated batch-processing
instrument capable of processing 40 samples per hour, with a
throughput capacity of 324 samples per 8-hour run. An integrated stirrer mixed the specimen briefly to disperse mucus
and aggregates. The specimen was then aspirated up into the
hollow stirrer, and dual-flow technology captured a representative sample on a frit-backed filter (MonoGen). The
MonoPrep filter was then pressed against the slide to transfer
the cells onto a 20-mm-diameter circular area ❚Image 2❚. After
cell transfer, the Processor applied a premeasured amount of
reagent alcohol fixative directly onto the slide, which was
allowed to dry. Slides were held in racks until stained. (They
can be held up to 7 days at room temperature to 37°C without
compromising specimen quality.)
© American Society for Clinical Pathology
Anatomic Pathology / ORIGINAL ARTICLE
Pap smears were screened and interpreted by the study
cytotechnologists and cytopathologists in the same manner as
their routine practice. All slides were interpreted and results
reported in accordance with the Bethesda System 2001 criteria and nomenclature.5 Specifically, the Bethesda System
2001 criteria for liquid-based preparations were used to determine the adequacy of the MonoPrep preparations. As for all
liquid-based preparations, a MonoPrep slide was required to
have a minimum of 5,000 well-visualized and well-preserved
squamous cells. Because the diameter of a MonoPrep cell
deposition circle is the same as that for a ThinPrep, the guidelines for estimating cellularity were the same as those for the
ThinPrep.5 An endocervical/squamous metaplastic component was not required for adequacy, but its presence or
absence was documented for all slides. For data analysis, all
cases were classified into 1 of 9 general diagnostic categories:
unsatisfactory; negative for an intraepithelial lesion or malignancy (NILM); NILM with reactive or reparative changes
(NILM-R); ASC-US; atypical squamous cells, cannot exclude
HSIL (ASC-H); atypical glandular cells (AGC); LSIL; highgrade squamous intraepithelial lesion (HSIL); and cancer.
Cancer was further subdivided into adenocarcinoma in situ
(AIS), squamous cell carcinoma (SCC), and adenocarcinoma
(AC). Representative images of MonoPrep cases are shown in
❚Image 3❚.
A screening cytotechnologist performed all first-pass
interpretations of conventional and MonoPrep specimens. Pap
smear slides were reviewed in the ordinary course of the cytologist’s workload with other nonstudy slides, including other
liquid-based Pap tests. On a given day, each cytotechnologist
was assigned study-related smears or MonoPrep slides but not
both. The slide allocation method was designed to ensure that,
at the end of the study, each cytotechnologist had read approximately the same number of MonoPrep slides and study-related smears and that they were masked as to matching pairs. It
was possible that a cytotechnologist would coincidentally read
both slides in a pair by random allocation, but not on the same
day, and only in a masked manner.
The designated QC cytotechnologist conducted a QC
review of all cases designated as unsatisfactory by the screening cytotechnologist and at least 10% of all cases interpreted
as NILM, including those from essentially all subjects meeting protocol-defined high-risk case criteria. At some laboratories, QC review of NILM and unsatisfactory slides was performed by the study cytopathologist. Cytotechnologists and
cytopathologists completed the study CRFs as they evaluated
each slide and had no further access to them.
The study coordinator at each laboratory oversaw daily
activities of the study at that site. This person was responsible
for maintaining the site notebook and ensuring prompt transmittal of CRF pages and resolution of data queries. All sites
were independently monitored and audited.
❚Image 1❚❚ The MonoPrep Processor uses robotic technology
to apply cells to a slide and bar code technology to ensure
positive specimen identification (maximum throughput of 324
specimens per 8-hour shift). After completing each cassette
of 40 slides, the Processor discharges the cassette into a
slide output bin on the right side of the instrument.
❚Image 2❚❚ Macroscopic view
of a MonoPrep slide.
If the MPPT or the smear slide from a case was interpreted as ASC-US or a more severe abnormality by the laboratory,
both slides were submitted for masked review by an independent pathologist (IP). In addition, if the MPPT or the smear were
interpreted as NILM-R, then, in virtually all (>98%) cases,
both slides were reviewed by an IP. Approximately 6% of cases
interpreted as NILM on both slides were submitted to an IP for
a masked reference interpretation. In contrast with prior splitsample studies for liquid-based Pap tests, cases were not
excluded from IP referral or performance analysis owing to a
laboratory interpretation of unsatisfactory on one slide. The
Am J Clin Pathol 2008;129:193-201
© American Society for Clinical Pathology
195
DOI: 10.1309/E63PQJJXWCDLWNHQ
195
195
Cibas et al / MONOPREP PAP TEST CLINICAL TRIAL RESULTS
A
B
C
D
❚Image 3❚❚ MonoPrep cases. A, Normal MonoPrep Pap slide showing numerous squamous cells. Large sheets of cells like the
one on the right are common on MonoPrep slides (Papanicolaou, ×100). B, Normal MonoPrep Pap slide showing benign
squamous and endocervical cells (Papanicolaou, ×400). C, Low-grade squamous intraepithelial lesion (Papanicolaou, ×400). D,
High-grade squamous intraepithelial lesion (Papanicolaou, ×400).
sets of MPPT and smear slides requiring IP review were independently and randomly allocated to 1 of the 5 board-certified
cytopathologists (R.M.A., D.R.B., M.D.G., M.R.H., and
A.T.M.) on the IP panel. Thus, each of the 5 IPs reviewed 20%
of the MPPT slides and smears.
For essentially all (>98%) cases submitted for IP review,
a test for HPV was performed using the hc2 HPV test (Digene,
Gaithersburg, MD) using the high-risk (hr) probe set. The test
was performed on 2 samples: a specimen co-collected (at the
time the smear was obtained) with a Digene cervical sampler;
and an aliquot from the MPPT specimen. HPV testing was
performed for 2 reasons: (1) to query the relative sensitivities
of the MPPT and smear for detection of hrHPV+ ASC-US,
196
196
Am J Clin Pathol 2008;129:193-201
DOI: 10.1309/E63PQJJXWCDLWNHQ
and (2) to validate the HPV assay on MPPT specimens.
Results from this part of the study will be reported separately.
For each case reviewed by an IP, the reference diagnosis
for the case was the most abnormal diagnosis from the 2 IPreviewed slides. Severity ranking was as follows: unsatisfactory < NILM < NILM-R < ASC-US (hrHPV–) < ASC-US
(hrHPV+) < ASC-H < AGC < LSIL < HSIL < AIS < SCC <
AC. The reference diagnosis was used as the cytologic “truth”
diagnosis for the case and was termed the reference diagnosis
by IP (RDIP). To assess the performance of the MPPT relative
to the conventional Pap smear for each IP-reviewed case, the
laboratory diagnoses made by the study sites using the 2 methods were compared with the RDIP.
© American Society for Clinical Pathology
Anatomic Pathology / ORIGINAL ARTICLE
Data management was performed by HHI Clinical
Research (Hunt Valley, MD). Completed CRF pages were
transmitted by the laboratory site to the data management
group via encrypted data fax, direct point-to-point fax, or by
courier in batches. The data in the CRF were entered independently by 2 data entry personnel into a database. If their
entries were identical, the entry was accepted into the data
record; if there was a mismatch, the difference was resolved
(eg, by examining the transcription or contacting the site for a
CRF correction in case of illegibility). All data also received
automated validity checks. Errors, ambiguity, or illegibility in
the data resulted in immediate query by the project manager to
the laboratory and/or collection site for prompt resolution.
Data listings and SAS tables (SAS Institute, Cary, NC)
were prepared for statistical analysis. The ASC-US/SIL ratio
was calculated by dividing the number of laboratory interpretations of ASC-US by the sum of LSIL + HSIL + AIS + SCC
+ AC interpretations. For analysis, the IP-established diagnosis was considered the subject’s “true diagnosis” (= RDIP) for
the calculation of true-positive and false-positive detection
rates. This true diagnosis was the more severe of the 2 interpretations made by the IP for the subject’s smear and MPPT
slide pair-mate. Laboratory true-positive and false-positive
detection rates for MPPTs and smears were calculated and the
95% confidence intervals (CIs) of their ratios evaluated for
subjects with the RDIP equal to the following cutoffs: ASCUS+, ASC-H/AGC+, LSIL+, HSIL+, and cancer.
Among the cases at or above an RDIP cutoff (eg, LSIL+),
a laboratory assessment for the MPPT or smear slide was considered a true-positive detection if the laboratory interpretation was at or above the RDIP cutoff (eg, laboratory assessment for RDIP = LSIL+ was LSIL, HSIL, or cancer).
Conversely, the laboratory diagnosis for the smear or MPPT
slide was considered a false-positive when the laboratory
interpretation was greater than the RDIP for the subject (eg, a
laboratory interpretation of HSIL was a false-positive if the
MPPT and smear are interpreted as LSIL or lower by the IP).
In determining true-positive and false-positive detection, the
IP interpretation for a slide did not need to match the laboratory interpretation for the same slide. In other words, a laboratory detection of an abnormality was considered a true
detection even when that abnormality was observed only on
the other slide on IP review. The ratios of true-positive rates
and ratios of false-positive rates and their 95% CIs were calculated for the cases with an RDIP of ASC-US+, ASCH/AGC+, LSIL+, HSIL+, and cancer.6 The statistical significance of ratios differing from 1.0 was demonstrated when the
95% CI did not include 1.00.7,8
The ability of each method to generate a satisfactory specimen was presented in a 2 × 2 table of satisfactory and unsatisfactory samples. A 2-sided McNemar test was conducted to
examine differences in adequacy. For this analysis, subjects
with unsatisfactory samples owing to collection site failure to
provide documentary information were excluded from both
arms. For the MonoPrep arm, a specimen was deemed unsatisfactory only if no satisfactory slide could be prepared from
the specimen.
The number and percentage of infectious agents, benign
conditions, and presence of a sufficient endocervical/transformation zone component (ECTZ) were summarized for the
MonoPrep and Pap smear methods. Two-sided McNemar tests
were used to examine rate differences between the methods.
Results
Paired specimens were obtained from 11,244 subjects, of
which 339 were received after the study cutoff date and were
not included in the analysis. Of the 10,905 accepted cases,
10,739 (98.5%) fulfilled eligibility criteria and were included
in the statistical analysis. A summary of subject demographics is displayed in ❚Table 1❚. The mean age was 35.4 years
(range, 18-90 years). A total of 3,500 (32.6%) subjects fulfilled criteria that placed them at high risk for developing cervical cancer; 509 (4.7%) were pregnant, and 1,402 (13.1%)
were postmenopausal.
The laboratory interpretations for the 10,739 smear and
MPPT pairs are displayed in ❚Table 2❚. The MPPT and smear
interpretations were exactly concordant in 8,290 cases
(77.2%). There are 1,404 cases in the lower left of Table 2 that
represent cases with a more severe interpretation on the MPPT
slide than on the smear pair-mate. This number is 34.4%
greater than the 1,045 cases in the upper right, which represent
cases with a more severe interpretation on the smear as compared with the MPPT. The ASC-US/SIL ratio was 1.77 for the
smears and 1.51 for the MPPT (–15%).
All 3,185 eligible cases were reviewed by an IP and generated an RDIP. The interpretations by the IP are displayed in
❚Table 3❚. There were 46 cases (MPPT, 36; smear, 23) with an
❚Table 1❚
Demographics for 10,739 Subject
Variable
No. (%)
US subjects
International subjects
Age (y)
Mean ± SD
Range
Ethnicity
White
Hispanic
Black
Other (or not provided)
Asian
Native American
Pacific
7,689 (71.6)
3,050 (28.4)
35.4 ± 12.2
18-90
5,213 (48.5)
2,690 (25.0)
1,400 (13.0)
1,141 (10.6)
227 (2.1)
37 (0.3)
31 (0.3)
Am J Clin Pathol 2008;129:193-201
© American Society for Clinical Pathology
197
DOI: 10.1309/E63PQJJXWCDLWNHQ
197
197
Cibas et al / MONOPREP PAP TEST CLINICAL TRIAL RESULTS
❚Table 2❚
Laboratory Interpretations for MPPTs and Smears (All Sites)
Smear Diagnosis
MPPT
Diagnosis
UNSAT
NILM
NILM-R
ASC-US
ASC-H
AGC
LSIL
HSIL
AIS
SCC
AC
Total
UNSAT
NILM
NILM-R
ASC-US
ASC-H
AGC
LSIL
HSIL
AIS
SCC
AC
Total
43
209
11
23
1
4
6
2
1
2
0
302
58
7,744
214
538
9
21
135
4
0
0
0
8,723
6
198
59
41
0
1
1
0
0
0
0
306
12
459
40
201
10
4
112
10
0
1
0
849
0
16
1
4
0
1
1
7
0
4
0
34
0
35
1
7
0
1
0
1
0
0
0
45
5
55
6
73
2
1
176
22
0
0
0
340
2
15
2
7
2
0
27
50
0
5
1
111
0
0
0
0
0
1
0
0
2
0
0
3
0
0
1
0
1
0
1
6
0
13
2
24
0
1
0
0
0
0
0
0
0
0
1
2
126
8,732
335
894
25
34
459
102
3
25
4
10,739
AC, adenocarcinoma; AGC, atypical glandular cells; AIS, adenocarcinoma in situ; ASC-H, atypical squamous cells, cannot exclude HSIL; ASC-US, atypical squamous cells of
undetermined significance; HSIL, high-grade squamous intraepithelial lesion or malignancy; LSIL, low-grade squamous intraepithelial lesion; MPPT, MonoPrep Papanicolaou
test; NILM, negative for intraepithelial lesion or malignancy; NILM-R, NILM with reactive and/or reparative changes; SCC, squamous cell carcinoma; UNSAT, unsatisfactory.
RDIP of cancer (AC, SCC, or AIS); 282 cases (MPPT; 223;
smear, 140) with an RDIP of HSIL; 609 cases (MPPT, 494;
smear, 310) with an RDIP of LSIL; 33 cases (MPPT, 25; smear,
12) with an RDIP of AGC; 131 cases (MPPT, 89; smear, 82)
with an RDIP of ASC-H; 801 cases (MPPT, 633; smear, 579)
with an RDIP of ASC-US; and 565 cases (MPPT, 528; smear,
449) with an RDIP of NILM-R. IP interpretations of an epithelial abnormality were significantly greater with the MPPT compared with the smear at every threshold (ASC-US+, 30.9%;
ASC-H/AGC+, 52.9%; LSIL+, 59.1%; HSIL+, 59%; and cancer, 57%) ❚Table 4❚.
True-positive and false-positive ratios for the MPPT and
smear interpretations at 5 diagnostic thresholds are shown in
❚Table 5❚. For the cases with an RDIP of ASC-US+, the MPPT
method detected 1.15 (1,274/1,111) times more true-positive
cases than the smear method for all sites combined. This
increase is statistically significant. The observed ratios of the
true-positive rates varied among the sites from 1.02 to 1.49.
The ratio of the false-positive rates for ASC-US+ was 0.92
(271/296) for all sites combined (range, 0.55-1.51). The
observed decrease in the false-positive MPPT rate relative to
the false-positive smear rate is not statistically significant.
For cases with an RDIP of ASC-H/AGC+, the MPPT
method detected 1.23 (537/438) times more true-positive
cases than the smear method detected for all sites combined.
This increase is statistically significant. The observed ratios
of the true-positive rates varied among the sites from 1.11 to
1.67. The ratio of the false-positive rates for ASC-H/AGC+
was 0.95 (115/121) for all sites combined (range, 0.631.35). The observed decrease in the false-positive MPPT
rate relative to the false-positive smear rate is not statistically significant.
❚Table 3❚
Independent Pathologist Interpretations for MPPT and Smears (All Sites)
Smear Diagnosis
MPPT
Diagnosis
UNSAT
NILM
NILM-R
ASC-US
ASC-H
AGC
LSIL
HSIL
AIS
SCC
AC
Total
UNSAT
NILM
NILM-R
ASC-US
ASC-H
AGC
LSIL
HSIL
AIS
SCC
AC
Total
26
100
62
67
11
1
35
8
1
7
0
318
24
568
217
248
27
13
136
38
1
0
0
1,272
8
174
104
89
18
3
34
18
0
1
0
449
11
162
93
131
12
3
116
50
0
1
0
579
4
17
14
22
6
2
6
8
0
2
1
82
1
3
4
2
1
0
0
1
0
0
0
12
5
36
23
56
8
1
153
28
0
0
0
310
3
14
11
17
6
1
13
66
0
9
0
140
0
0
0
0
0
0
0
1
0
0
1
2
0
0
0
0
0
1
1
5
0
10
0
17
1
0
0
1
0
0
0
0
0
1
1
4
83
1,074
528
633
89
25
494
223
2
31
3
3,185
AC, adenocarcinoma; AGC, atypical glandular cells; AIS, adenocarcinoma in situ; ASC-H, atypical squamous cells, cannot exclude HSIL; ASC-US, atypical squamous cells of
undetermined significance; HSIL, high-grade squamous intraepithelial lesion or malignancy; LSIL, low-grade squamous intraepithelial lesion; MPPT, MonoPrep Papanicolaou
test; NILM, negative for intraepithelial lesion or malignancy; NILM-R, NILM with reactive and/or reparative changes; SCC, squamous cell carcinoma; UNSAT, unsatisfactory.
198
198
Am J Clin Pathol 2008;129:193-201
DOI: 10.1309/E63PQJJXWCDLWNHQ
© American Society for Clinical Pathology
Anatomic Pathology / ORIGINAL ARTICLE
❚Table 4❚
2 × 2 Presentations of Independent Pathologists’ Interpretations
MonoPrep Diagnosis
ASC-US+
<ASC-US
≥ASC-US
Total
ASC-H/AGC+
<ASC-H/AGC
≥ASC- H/AGC
Total
LSIL+
<LSIL
≥LSIL
Total
HSIL+
<HSIL
≥HSIL
Total
Cancer†
<Cancer
≥Cancer
Total
Glandular abnormalities
Other‡
AGC/AIS/adenocarcinoma
Total
Smear Diagnosis
Total
<ASC-US
1,283
756
2,039
≥ASC-US
402
744
1,146
<ASC-H/AGC
2,084
534
2,618
≥ASC- H/AGC
234
333
567
2,318
867
3,185
<LSIL
2,248
464
2,712
≥LSIL
184
289
473
2,432
753
3,185
<HSIL
2,857
165
3,022
≥HSIL
69
94
163
2,926
259
3,185
<Cancer
3,139
23
3,162
≥Cancer
10
13
23
3,149
36
3,185
Other‡
3,139
28
3,167
AGC/AIS/adenocarcinoma
16
2
18
3,155
30
3,185
Ratio (95% CI)
P*
1.31 (1.24-1.38)
.00001
1.53 (1.42-1.65)
.00001
1.59 (1.46-1.73)
.00001
1.59 (1.37-1.84)
.00001
1.57 (1.06-2.31)
.024
1.67 (0.95-2.92)
.070
1,685
1,500
3,185
AGC, atypical glandular cells; AIS, adenocarcinoma in situ; ASC-H, atypical squamous cells, cannot exclude HSIL; ASC-US, atypical squamous cells of undetermined
significance; CI, confidence interval; HSIL, high-grade squamous intraepithelial lesion or malignancy; LSIL, low-grade squamous intraepithelial lesion.
* McNemar test.
† Includes squamous cell carcinoma, adenocarcinoma, and AIS.
‡ Includes all categories except AGC, AIS, and adenocarcinoma.
For cases with an RDIP of LSIL+, the MPPT method
detected 1.26 (486/385) times more true-positive cases than
the smear method detected for all sites combined. This
increase is statistically significant. The observed ratios of the
positive rates varied among the sites from 1.14 to 1.53. The
ratio of the false-positive rates for LSIL+ was 1.13 (107/95)
for all sites combined (range, 0.71-1.56). The observed
increase in the false-positive MPPT rate relative to the falsepositive smear rate is not statistically significant.
For cases with an RDIP of HSIL+, the MPPT method
detected 1.04 (113/109) times more true-positive cases than
the smear method detected for all sites combined. This
increase is not statistically significant. The observed ratios of
the true-positive rates varied among the sites from 0.96 to
1.33. The ratio of the false-positive rates for HSIL+ was 0.68
(21/31) for all sites combined (range, 0.40-1.33). The
observed decrease in the false-positive MPPT rate relative to
the false-positive smear rate is not statistically significant.
❚Table 5❚
True- and False-Positive Results for MPPT and Smear as Measured Against a Reference Diagnosis by an IP
Positive by
Diagnostic
Threshold
Positive
by IP
MPPT
ASC-US+
ASC-H/AGC+
LSIL+
HSIL+
Cancer*
1,902
1,101
937
328
46
1,274
537
486
113
28
Positive by
Smear
TPRMPPT/TPRsmear
Ratio (95% CI)
Not Positive
by IP
MPPT
Smear
FPRMPPT/FPRsmear
Ratio (95% CI)
1,111
438
385
109
23
1.15 (1.09-1.21)
1.23 (1.14-1.32)
1.26 (1.17-1.36)
1.04 (0.89-1.20)
1.22 (0.87-1.70)
1,283
2,084
2,248
2,857
3,139
271
115
107
21
4
296
121
95
31
6
0.92 (0.78-1.07)
0.95 (0.75-1.21)
1.13 (0.87-1.45)
0.68 (0.41-1.13)
0.67 (0.25-1.78)
AGC, atypical glandular cells; ASC-H, atypical squamous cells, cannot exclude HSIL; ASC-US, atypical squamous cells of undetermined significance; CI, confidence interval;
FPR, false-positive rate; HSIL, high-grade squamous intraepithelial lesion or malignancy; IP, independent pathologist; LSIL, low-grade squamous intraepithelial lesion; MPPT,
MonoPrep Papanicolaou test; TPR, true-positive rate.
* Includes squamous cell carcinoma, adenocarcinoma, and adenocarcinoma in situ.
Am J Clin Pathol 2008;129:193-201
© American Society for Clinical Pathology
199
DOI: 10.1309/E63PQJJXWCDLWNHQ
199
199
Cibas et al / MONOPREP PAP TEST CLINICAL TRIAL RESULTS
For the 46 cases with an RDIP of cancer, the MPPT
method detected 1.22 (28/23) times more true-positive cases
than the smear method detected for all sites combined. This
increase is not statistically significant and has a 95% CI of
0.87 to 1.70. The ratio of the false-positive rates for cancer
was 0.67 (4/6) for all sites combined (range, 0.5-0.75). The
observed decrease in the false-positive MPPT rate relative to
the false-positive smear rate is not statistically significant.
❚Table 6❚ shows a comparison of specimen adequacy
interpretations for the conventional smear and the MPPT
methods for all sites combined. There were 302 interpretations
of unsatisfactory on the smears and 126 (58% fewer) on the
MPPT preparations. The unsatisfactory rate was significantly
lower (P < .00001) for the MPPT (1.2%; range for all sites,
0.3-2.0%) than for the smear method (2.8%; range for all sites,
0.8-3.9).
Laboratories assessed slides for the presence of an adequate ECTZ. As shown in ❚Table 7❚, the ECTZ was absent or
insufficient in 43 fewer MPPT than smear slides (–3.3%; 95%
CI, –8.5% to 2.1%; P = .225), a difference that is not statistically significant. There were equivalent distributions in the
number of MPPTs and smears with ECTZ considered
“detectable” (1-10 cells), “typical” (11-25 cells), and “abundant” (>25 cells) (data not shown). Similar equivalence
between the MPPT and smear was observed with the distributions between MPPT and smear in the quantity of abnormal
cells (data not shown).
The MPPT vs smear interpretations showed no statistically significant difference in the detection of reactive or reparative conditions and infectious agents. ❚Table 8❚ shows the
detection rates for these conditions and agents.
❚Table 6❚
Specimen Adequacy by Laboratory Assessment
Smear
MonoPrep Test
Unsatisfactory
Satisfactory
Total
Unsatisfactory
Satisfactory
Total
43
259
302
83
10,354
10,437
126
10,613
10,739
❚Table 7❚
Endocervical/Transformation Zone Component
Smear
MonoPrep Test
Absent
Detectable
Total
Absent
Detectable
Total
640
649
1,289
606
8,604
9,210
1,246
9,253
10,499
200
200
Am J Clin Pathol 2008;129:193-201
DOI: 10.1309/E63PQJJXWCDLWNHQ
❚Table 8❚
Benign Conditions*
Condition
MonoPrep Test
(n = 10,739)
Smear
(n = 10,739)
Reactive or reparative
Inflammation
Intrauterine device
Atrophic vaginitis
Radiation
Other†
Infectious agent
Candida/fungus
Trichomonas vaginalis
Actinomyces
Bacterial vaginosis (coccobacilli)
Herpes simplex
Other‡
335 (3.1)
249 (2.3)
0 (0.0)
0 (0.0)
3 (≤0.1)
83 (0.8)
1,507 (14.0)
523 (4.9)
105 (1.0)
0 (0.0)
980 (9.1)
3 (≤0.1)
0 (0.0)
306 (2.8)
231 (2.2)
4 (≤0.1)
1 (≤0.1)
1 (≤0.1)
77 (0.7)
1,496 (13.9)
426 (4.0)
158 (1.5)
0 (0.0)
1,035 (9.6)
9 (0.1)
2 (≤0.1)
*
Data are given as number (percentage). Totals are different from the sums of
categories because some cases had multiple assessments.
† Includes reactive conditions that were not otherwise specified and uncommon
observations such as chemical irritation, drug reactions, and cervical trauma.
‡ Includes appearance of microbial infection or sequela of unidentified or unusual
taxonomy.
Discussion
In this study, the MonoPrep Pap test was evaluated by comparing its performance with that of the direct smear, and relative
sensitivity and specificity were calculated based on an independent assessment of the slides by a reference pathologist.
This method of test comparison was used in the evaluations of
the ThinPrep and the SurePath Pap tests.2,9 This study has
shown that slides prepared by the MPPT, as compared with
smears, yield statistically significant increases in true-positive
cytologic results, as judged by an IP, for the diagnostic thresholds ASC-US+, ASC-H/AGC+, and LSIL+. Increases at these
thresholds were observed for all study laboratories. Based on
independent reference review interpretations, disease detection with MPPT was significantly increased at all thresholds
(ASC-US+, ASC-H/AGC+, LSIL+, HSIL+, and cancer).
Comparisons of false-positive rates did not show a statistically significant difference between MPPT and smears at these
(or any) diagnostic thresholds, confirming that the increase in
relative sensitivity achieved with the MPPT reflected a true
increase in diagnostic accuracy.
These results are arguably stronger than those obtained in
the multicenter split-sample clinical trials of 7,360 ThinPrep
Pap tests and 8,983 SurePath Pap tests.1-3,9 The data in those
reports and in this one are not directly comparable because of
different cohorts and a difference in the statistical analysis.
Nevertheless, in the SurePath (formerly AutoCyte PREP) clinical trial,3 the difference between smear and SurePath in the
detection of epithelial lesions was not significant for ASC-US
or LSIL+. In the ThinPrep clinical trial, the increase in detection was significant for ASC-US+ and LSIL+ only for a subset of trial sites.1
© American Society for Clinical Pathology
Anatomic Pathology / ORIGINAL ARTICLE
The inability to demonstrate a significant difference in the
detection of HSIL+ in any of the 3 clinical trials likely represents the limitations of sample size. It may also reflect, in this
study, the limitations of the study protocol, which defined
truth as an RDIP based on the highest abnormality observed
on either or both slides of a case. This formulation allows a
laboratory overcall of HSIL to be tallied as a correct interpretation even when it is not confirmed for that slide by IP review
(ie, the abnormality is correctly interpreted only on the other
slide). If one examines the IP interpretations independently
(Table 3) and accepts them as truth, then the MPPT shows a
significant improvement over the smear at the thresholds of
HSIL+ and cancer.
The method for assessing test accuracy in this study,
which was based on review of the slides by an IP, met the
approval of the FDA and was also used in the clinical trials for
the approvals of the ThinPrep2 and SurePath9 Pap tests. Data
using other accuracy assessment methods (correlation with
HPV and biopsy results) were also obtained; those results will
be published separately.
The MPPT is the first liquid-based Pap method approved
by FDA since 1999. MPPT slides show the uniformity and the
reduction of artifacts commonly associated with other liquidbased Pap tests. The MonoPrep Processor also incorporates
several enhancements. It is a fully automated, batch-processing instrument with end-to-end bar-coding technology that
ensures specimen integrity and chain of custody. It prevents
cross-contamination by automated vial and waste handling,
aerosol prevention, discrete slide fixation and dry slide storage, and automated vial resealing using plastic film. The
MPPT system is an effective method for preparing gynecologic cytology slides to screen for cervical abnormalities and can
serve as a replacement for the Pap smear.
From the Departments of 1Pathology, Brigham and Women’s
Hospital and Harvard Medical School, Boston, MA; 2Preventive
Medicine, University of Southern California, Arcadia; 3Pathology,
Medical University of South Carolina, Charleston; and
4AmeriPath Utah, Sandy; 5DCL Medical Laboratories,
Indianapolis, IN; 6Cleveland Clinic Florida Hospital, Naples;
Indiana, Indianapolis; 8CytoLaboratories, San
Antonio, TX; 9Pathology Services, Cambridge, MA; and
10Universal Diagnostic Laboratories, New York, NY,
7AmeriPath
Supported by MonoGen.
Address reprint requests to Dr Cibas: Dept of Pathology,
Brigham and Women’s Hospital, 75 Francis St, Boston, MA 02115.
Drs Bolick and Cibas were members of the Medical Advisory
Board of MonoGen at the time of the study.
References
1. Lee KR, Ashfaq R, Birdsong GG, et al. Comparison of
conventional Papanicolaou smears and a fluid-based, thinlayer system for cervical cancer screening. Obstet Gynecol.
1997;90:278-287.
2. Sherman ME, Mendoza M, Lee KR, et al. Performance of
liquid-based, thin-layer cervical cytology: correlation with
reference diagnoses and human papillomavirus testing. Mod
Pathol. 1998;11:837-843.
3. Bishop JW, Bigner SH, Colgan TJ, et al. Multicenter masked
evaluation of AutoCyte PREP thin layers with matched
conventional smears: including initial biopsy results. Acta
Cytol. 1998;42:189-197.
4. National Committee for Clinical Laboratory Standards.
Papanicolaou Technique; Approved Guideline. 2nd ed. Wayne,
PA: National Committee for Clinical Laboratory Standards;
2001. NCCLS document GP15-A2.
5. Solomon D, Davey D, Kurman R, et al. The 2001 Bethesda
System: terminology for reporting results of cervical cytology.
JAMA. 2002;287:2114-2119.
6. Schatzkin A, Connor RJ, Taylor PR, et al. Comparing new
and old screening tests when a reference procedure cannot be
performed on all screeners. Am J Epidemiol. 1987;125:672-678.
7. Newcombe RG. Interval estimation for the difference between
independent proportions: comparison of eleven methods. Stat
Med. 1998;17:873-890.
8. Cheng H, Macaluso M. Comparison of the accuracy of two
tests with a confirmatory procedure limited to positive results.
Epidemiology. 1997;8:104-106.
9. Hessling JJ, Raso DS, Schiffer B, et al. Effectiveness of thinlayer preparations vs conventional Pap smears in a blinded,
split sample study: extended cytologic evaluation. J Reprod
Med. 2001;46:880-886.
Am J Clin Pathol 2008;129:193-201
© American Society for Clinical Pathology
201
DOI: 10.1309/E63PQJJXWCDLWNHQ
201
201